The proposed research manipulates semantic relatedness (similarity) of words and examines its effects on recall and recognition. Its impacts on basic memory theories and issues of false memories will be examined. Semantic similarity in information is a major source of a type of memory errors called false (illusory) memory, to which senile and brain-damaged people are especially vulnerable. The nature of the false memory and the procedures to reduce/eliminate it are studied. The questions to be examined are: Does semantic relatedness affect recall and recognition in the same or opposite ways? Does it interact with memory-set size variable in a Steinberg task? Does it produce a mirror effect? Does it facilitate or inhibit recall output? At what memory-set size do semantically related intrusions and false positive recognitions occur? The answers to these questions will shed light on the nature of the memory processes underlying recall and recognition, and have important theoretical implications for several major memory theories, i.e., the Spreading Activation, Output Interference, Serial Memory Scan, Search of Associated Memory, Mirror Effects, and Relative List Strength theories. Memory errors from related and unrelated materials will be comparatively analyzed. True memory and false memory will be contrasted in recognition response time, word recall latency, word output serial position, and learning dissociation effects. Although false memory is phenomenally as realistic as true memory, the proposed research will demonstrate that it can be reliably distinguished from true memory by these measures. Additionally, this research will propose a theory and method for reducing/eliminating false memory in which people can be made aware of their initially creating false memory and internally editing it out. It will be shown that by undergoing this metacognitive self-correcting process, people can be made to avoid similar false memory on a later test. Finally, a theory and method of cumulative levels of processing will be proposed to explain why deeper processing of information cannot consistently reduce false memory and how the proposed method can more reliably reduce it. The findings from this study will contribute to efforts towards rehabilitating senile and brain-damaged people who are found to be especially prone to false memory.